KR20000048187A - Adjustment of call selection to achieve target values for interval-based performance metrics in a call center - Google Patents

Abstract

PURPOSE: An adjustment of call selection is provided for call centers or other call processing systems in which voice calls, e-mails, faxes, voice messages, text messages, Internet service requests or other types of communications are distributed among a number of service agents for handling. CONSTITUTION: A call selection process in a call center is controlled so as to ensure the achievement of target values for one or more interval-based performance metrics. In an illustrative embodiment, a memory in the call center is used to store information regarding contractual target values of one or more interval-based performance metrics such as, for example, an average speed of answering metric, or a percent in service level metric. The call selection process is then adjusted within a given performance interval based at least in part on a comparison of a value of the metric actually achieved during the interval to the corresponding stored target value. For example, a service objective of the call selection process may be adjusted at one or more designated points in the interval if the value of the metric actually achieved to a given one of the points will not allow achievement of the target value within the interval.

Description

An article of manufacture comprising a selection processing control method, a communication selection processing control device, and software.

The present invention generally relates to a call center or other call processing system for distributing a plurality of service agents to handle voice calls, e-mails, faxes, voice messages, text messages, Internet service requests or other forms of communication. .

Call centers distribute calls and other forms of communication to useful call-handling service agents in accordance with various predetermined standards. In current call centers, call center operators can often program standards for handling calls using a capability known as call vectoring. Typically, when a call center detects that an agent can handle a call usefully, the call center checks the agent's call-handling skill and, according to its normal priority, The agent forwards the longest-waiting call corresponding to the agent's highest-priority. Most conventional call distribution techniques generally focus on "fairing" the caller and the agent. This fairness is reflected for the slowest most-idle-agent call assignment process by standard first-in, first-out. The skills-based queuing mentioned above, based on this basic processing, allows each agent to be placed in a plurality of different service categories depending on the type of skill that the agent supports. Improve.

Today, many call centers are run by companies acting as service bureaus, which means that the service bureaus handle calls in contracts with other companies. The contract generally requires how the service bureau should handle the call, how much is paid for the service bureau to achieve performance metrics based on the specified intervals, and if the performance metrics do not match, the service bureau should contact the service bureau. Determine financial penalties. Designated performance metrics generally “fit” or “do not match” based on the overall run for a given time, such as for example half an hour or an hour. For example, the contract may specify an average speed of answer (ASA) of calls of type A within 30 seconds within a predetermined interval, which may result in some callers if the ASA is less than 30 seconds within the interval. Means they can wait longer. In another example, the contract starts with the percentage of calls that must be handled within a given time within a specified time interval, for example, at 80% of calls handled within 20 seconds at a specified interval. Percentage in service level (PSL) can be specified.

Lucent Technologies Corporation Current call centers, such as, have the ability to coordinate call selection processing to handle calls based on a predicted wait time (PWT) in comparison with the service object (SO) that handles the call. In call centers with this type of capability, each call is still generally considered separately, regardless of any overall execution statistics already accumulated within the current interval. In addition, service targets are generally static and typically only changed by an administrative process. However, in service bureau applications, agents can handle various proficiencies managed by various contract agreements. The disadvantage of current call centers is how well such call centers generally achieve contract target values at intervals for execution metrics based on intervals in call selection processing, and achieve the contract target values. It does not take into account the fines associated with failure to do so.

The present invention is to provide call selection based on the target value of the interval-based execution metrics, to ensure the achievement of the target values for a particular execution interval. In an example embodiment, memory in the call center is used to store information regarding contract target values of execution metrics based on one or more intervals. Execution metrics based on the interval can be, for example, the average rate of the response metric or a percentage of the service level metric. At this time, the call selection process is adjusted at least in part within the predetermined execution interval by comparing the stored target value with the value of the metric actually achieved during the elapsed time within the predetermined interval. For example, if the value of the metric actually achieved at a given point in the interval does not allow the achievement of a target value within that interval, then based on the number of predicted calls predicted in the remaining interval of that interval, You can adjust the service target of the call selection process at the specified point. In addition, the call center may determine a fine relating to a deviation from the contract target value and consider this fine in coordinating the call selection process. Other techniques may be used to provide coordination of call selection processing, such as, for example, adjusting overload thresholds that control when one or more reservation agents start handling calls.

Advantageously, in the present invention, service bureaus or other call centers may meet certain contract requirements with minimal over- and minimum non-execution, thereby maximizing the contractually realized benefits. Can be. These and other features and advantages of the present invention can be clearly understood from the accompanying drawings and the following detailed description.

1 is a block diagram of a call center incorporating an exemplary embodiment of the present invention.

2 is a block diagram of an automatic call distribution (ACD) system implemented in the call center of FIG.

3 is a flow diagram illustrating adjustment of call selection to achieve performance metrics based on intervals in accordance with the present invention.

4 is an exemplary diagram in which call selection is adjusted in accordance with the present invention to achieve performance metrics based on intervals.

* Description of the main symbols in the drawing

101: ACD system 112: interface

113: exchange 114: service circuit

115: memory 116: processor

120: call queue 130: agent queue

140: Call Vector 150: Agent Selector

Although the present invention is described below in connection with call processing in an exemplary call center, it is not limited to the use of certain types of call centers or communication processing applications. For example, the present invention can be applied to processing related to reception of communication, transmission of communication, or transmission and reception of communication. The techniques described include automatic call distribution (ACD) systems, telemarketing systems, private-branch exchange (PBX) systems, systems based on computer-telephone integration (CTI), and these and other forms of call centers. Can be used in combination. The term "call center" described in this document encompasses any form of ACD system or telemarketing system and also handles calls, requests for other services, such as voice calls, video calls, multimedia calls, e-mail. Terminology encompassing other communication systems, handling various combinations of faxes or voice messages, and these and other forms of communication. The term “interval based performance metric” as described herein encompasses any measuring device or other indicator that characterizes the execution of at least one function of a call center, for example, during a particular time period. Examples of performance metrics based on intervals are the average answer rate (ASA) and the percentage of service level (PSL).

1 illustrates an exemplary call center in which the present invention may be implemented. The call center includes a plurality of telephone lines and / or trunks 100 that are selectively interconnected with a plurality of agent locations 102-104 through the ACD system 101. Each agent location 102-104 includes a voice and data terminal 105 that the corresponding agent 106-108 uses in handling the call. The terminal 105 is connected to the ACD system 101 via a voice and data transmission medium 109. The ACD system 101 includes a conventional basic call management system (BCMS) 110 and is connected to a conventional external call management system (CMS) 111. BCMS 110 and CMS 111 collect call records and call center statistics, as well as other information for use in managing the call center, collecting the call center records, and executing other functions. In alternative embodiments, the functions of the BCMS 110 and the CMS 111 may be provided inside or outside the ACD system 101 using a single call management system.

The ACD system 101 is for example Definity, a product of Lucent Technologies, Inc. It can be implemented in a manner similar to that of a PBX-based ACD system. 2 shows a simplified block diagram of one implementation that may be implemented in ACD system 101. The system 101 as shown in FIG. 2 is a stored-program-controlled system and has an interface 112 to an external communication link, a communication exchange 113 and a service circuit 114. (E.g., voice generator, announcement circuit, etc.), a memory 115 for storing control programs and data, and a stored control program to control the interface and switch and provide automatic call distribution. Processor 116 (eg, microprocessor, CPU, computer, etc.).

Referring again to FIG. 1, exemplary data elements stored in memory 115 of ACD system 101 include a series of call queues 120 and a series of agent queues 130. Each call queue 121-129 of the series of call queues 120 corresponds to a different agent skill level, and each agent queue 131-139 of the series of agent queues 130 is likewise. Corresponds to different agent proficiency. As in conventional systems, calls are prioritized, for example, calls may be queued to their priority in each of the call queues 120, or different call queues among a plurality of call queues. Can be queued such that each call queue corresponds to a different priority while corresponding to skill level. Similarly, the proficiency of each agent is prioritized according to its level of expertise in that skill, which agents are queued to that level of expertise in each queue of the agent queue 130, for example. Alternatively, in different agent queues of the plurality of agent queues, the agent queues may correspond to the proficiency while each of the agent queues corresponds to a different level of expertise. It is to be understood that the present invention may also be used in systems utilizing various other types of queue apparatuses and queue techniques.

The ACD system 101 also includes a call vector 140. The call vector 140 may be one of various types of storage control programs executed in the system 101. Calls entering the call center via link or trunk 100 are assigned to respective call queues 121-129 by call vector 140, based on the agent skills required to properly handle the call. Agents 106 to 108 useful for handling calls are assigned to agent queues 131 to 139, based on the proficiency they possess. An agent can have various proficiencies, and thus can be assigned to multiple agent proficiencies 131-139 simultaneously. Such agents are referred to herein as "multi-skilled agents". In addition, because an agent may have varying levels of expertise (eg, different levels of proficiency in multi-level grades, or first and second grade levels), different expertise Levels may be assigned to different agent queues 131-139. For call vectoring, published on November 3, 1993, AT & T Publication No. 555-230-520.

It is described in detail in the above, which is incorporated herein by reference. ACD techniques based on proficiency are described in detail, for example, in US Pat. No. 5,206,903, which is incorporated herein by reference.

In the ACD system 101, another program is executed using the agent selector 150. Agent selector 150 may read memory 115 of system 101, peripheral memory (eg, disk, CD-ROM, etc.) of system 101, or other form of computer reading associated with system 101. It may be executed by software stored on a possible medium, and may also be executed by the processor 116 or other suitable process hardware associated with the ACD system 101. The selector 150 in this example embodiment implements conventional techniques to provide assignment between useful calls and useful agents. Conventional techniques implemented by the selector 150 are well known in the art and are not described herein any further. It is to be understood that these functions may be executed on other elements of system 101 and may also be performed using a combination of other elements in such a system. A detailed description of call processing implemented in such a system, such as the ACD system 101, is filed, for example, on March 7, 1997, and entitled "Pending Call Selection Based on Predicted Wait Time", US Patent No. 08 / 813,513 and Feb. 12, 1998, described in US Patent No. 09 / 022,959, entitled "Call Center Agent Selection to Optimize Call Waiting Time," which is incorporated by reference. It is incorporated herein by reference.

In accordance with the present invention, the call center of FIG. 1 has the ability to coordinate call selection processing to achieve target values for interval based execution metrics. The call selection process that is adjusted need not be in any particular form in order to benefit from the techniques of the present invention. In a conventional call center, if a multi-skill agent is useful, then a "greatest need" form of call selection processing is performed at the head of the queues for proficiency that the multi-skill agent can handle. One call may be selected by examining the predicted wait times (PWTs) of the two, and then comparing the PWTs of the calls with the service subject (SO) identified for the proficiency. In an exemplary embodiment of the present invention, for a service bureau that must meet a contractually specified target value in an execution metric, such as an average response rate (ASA), the SO must be initially established at the ASA. Therefore, a call with PWT / SO = 1 must have an actual ASA that exactly corresponds to the target ASA specified in the contract, whereas a call with PWT / SO> 1 causes the actual ASA to be above the specified target value. do. As pointed out above, in these and conventional systems, service objects are generally static and only changed through a management process.

The present invention provides that the service bureau goes beyond the target value, i.e., over-performance, or beyond the target value, i.e., under the target, rather than simply handling, for example, a "maximum need" call at any particular moment. Recognize that it is desirable to obtain an ASA or other performance metric as close to the contractual target as possible without performance. The "maximum need" call selection process mentioned above is generally defined by the needs of a particular call with respect to other calls. The service bureau may not want to select the call with the maximum need, but instead may want to select the call that will make the most necessary contribution to achieving the contractual target value for the execution metric based on the interval. Using this contractual execution category actually makes a call selection decision that is completely different from the conventional "maximum need" selection. For example, if the service level for a particular proficiency is poor at the current interval, a very fast response to a series of calls becomes very important. In situations where the performance target value is at stake, in order to allocate more resources to proficiency, acceptable service levels may be adjusted periodically, as described in detail later.

3 is a flowchart illustrating a method of executing call selection to ensure the achievement of target values for performance metrics based on intervals in accordance with the present invention. In an example embodiment, it is assumed that one or more functions associated with the flow diagrams properly and unlimitedly execute program instructions calculated and stored by the processor 116 of the ACD 101 operating in combination with the memory 115. In step 200, execution intervals or intervals associated with a particular contract are identified along with the execution metric target values that must be achieved during that interval. Accordingly, determination of metric target values is made for one or more contracts, and various different call types and proficiency. In this step, the penalty for deviating from each contractual target value is defined in common units and under which condition (s) the penalty is incorporated into the decision regarding the adjustment of the subject of service.

In step 210, at a given point within a given execution interval, the actual achieved value is calculated for the execution metric as well as the number of predicted calls remaining in the interval. The predicted number of calls is, for example, the percentage of calls per minute that have been received so far in time for the remaining minutes in the interval, and the number of calls received so far from the estimate of the number of calls predicted for the entire interval. It can be calculated as In step 220, a new service target is calculated based on the actual achieved value and the predicted remaining number of calls. The new service target is a service target designed to ensure that a target value for an execution metric is achieved in the interval. In step 230, the calculation of steps 210 and 220 is repeated at one or more other designated points in the interval. As shown in step 240, after completion of the current interval, the process may be restarted for another interval. So steps 210-230 can be repeated every series of intervals, for which performance metrics must be observed.

As an example, assume that the contractual ASA target for calls requiring a particular skill is 30 seconds for a 30 minute interval. If 20 minutes have elapsed during the 30-minute interval and the actual ASA is 35 seconds, there will be only one third of the remaining call volume left by retracting the actual ASA toward the 30-second target value. Assuming that the amount of calls for the 30 minute interval is 100 calls and the number of calls handled up to this point is also 67, the 33 remaining calls predicted during that interval are combined within exactly 20 seconds to achieve the 30 second interval target. Must have an ASA. Therefore, acceptable service levels for proficiency can be readjusted using the system for, for example, about 18 seconds in the remainder of the interval. Similarly, if the actual ASA achieved in the first 67 calls in the first 20 minutes of the 30 minute interval was 20 seconds, then the 33 remaining calls predicted in that interval would have an ASA of about 50 seconds without jeopardizing the ASA target for the entire interval. Can be.

This type of adjustment may also be made in embodiments where the contractual target value is 80% of the calls handled within 20 seconds, with respect to the percentage in the service level (PSL) metric. If the actual achieved value for the PSL metric progressing in the interval causes the target value to deviate, the system may respond by running at a level above the target value PSL for the remaining intervals of the interval. Similarly, if the actual PSL progressing in that interval is much better than the contractual target value for that interval, the system can run at a level below the target value PSL for the remaining intervals of that interval. In other embodiments, multiple performance metrics may be observed and adjusted. For example, one embodiment may be configured that utilizes both ASA and PLS as performance metrics. In such a case, adjustment of call selection is possible to control the other one even if one of the execution metrics is at an acceptable level.

At the beginning of each new execution interval, it is possible to reinitialize the actual values of the service objects and the execution metrics, and to restart the calculation process and the adjustment process. Thus, the service object can be adjusted periodically over a predetermined interval. For example, if the interval is 30 minutes, the service object can be readjusted at designated points corresponding to 10, 20, and 25 minutes for that interval. For an interval of 60 minutes, the service object can be adjusted again at 15, 30, 45, 50, and 55 minutes for that interval. The period of coordination may be fixed or variable. In one alternative possible embodiment, for example, adjustments can be made continuously for the interval at which each call is handled. Of course, various other cycles, techniques, and categories may be used. In addition, as mentioned above, the process of adjusting the service subject may be affected by other factors such as fines for exceeding one or more target values and other contractual executions. For example, if it is more important to achieve a certain contract target value than to achieve one or more other contract target values, at a further expense, the service subject may be turned toward achieving the predetermined contract target value. In addition, adjustment of call selection to achieve the contract subject can be made, for example, by adjusting overload thresholds, so that reservation agents may respond earlier or later than they did not make the adjustment. Can be.

4 is a detailed diagram illustrating in more detail the adjustment of the service object SO to achieve target values for performance metrics based on intervals in an exemplary embodiment of the present invention. In this embodiment, the execution interval is 30 seconds, and the execution metric based on the interval is an ASA with a contract target value of 20 seconds. The embodiment is shown in a table, in which the first column represents the elapsed time in minutes, the second column represents the current SO, and the third column represents the number of calls handled in the interval so far. Column 4 represents the actual ASA achieved in the intervals so far, column 5 represents the predicted number of calls in the remaining intervals of the interval, and column 6 represents the new SO adjusted in accordance with the techniques of the present invention. Indicates.

At the time 10 minutes elapsed in this interval, the current SO is 20 seconds, the target value ASA, and 47 calls were handled on the actual ASA of 24 seconds. The predicted number of calls in the remaining interval is 94. The new SO is then calculated to achieve a target value ASA based on the actual ASA and the number of calls in the remaining intervals of the interval. As shown in FIG. 4, the calculation at line 1 in the table yields a new SO of 18 seconds. Therefore, the SO is adjusted downward to reflect the fact that the actual ASA is not better than the 20 second target value.

At the time 20 minutes elapsed in the interval, the current SO is 18 seconds, the new SO calculated on the previous line, and 90 calls were handled on the actual ASA of 21 seconds. The number of calls in the remaining interval is 45. According to the scheme shown in FIG. 4, the new SO is recalculated and 18 seconds of SO is rebuilt. The SO remains adjusted downward because the actual ASA is still worse than the 20 second target value. At the time 25 minutes elapsed in this interval, the current SO is 18 seconds, the SO calculated on the previous line, and 108 calls were handled in the actual ASA of 18 seconds. The number of calls in the remaining interval is 22. As shown, the new SO reflects the fact that the actual ASA is now better than the target value ASA. Therefore, the new SO is adjusted upward for the calculated value of 29 seconds. This value provides 18 actual ASAs at the completion of the 30 second interval, which corresponds to the contract target value in the performance metric.

Similarly, if an interval-based contract execution metric is specified in terms of percentages in the service level (PSL), then the call selection service target can be used to achieve the target value by comparing the actual PSL obtained at various points with the target value over that interval. Adjust at. Advantageously, in the present invention, a service bureau or other call center operator can meet both one or more contract requirements with both minimal over-execution and minimal non-execution, whereby a given contract or set of contracts can be met. The realized gain can be maximized.

The invention is particularly suited for use in call centers acting as service bureaus, but only partially acting as service bureaus, ie call centers using call centers for operations of other companies as well as their own operations. It is also applicable to call centers that operate similarly to the operation of the service bureau for various office devices within a single company. In the latter case, agreements can be made with office devices to control the level of service provided as in-home resources for employees and / or resources for customers.

The call selection function according to the present invention may be implemented using one or more elements of ACD system 101, such as, for example, agent selector 150. More generally, call processing to achieve interval based execution metrics may be executed by the processor 116 shown in FIG. 2 using program instructions and information stored in memory 115. In another embodiment of the present invention, other elements shown in FIG. 1, namely call centers or other forms of call centers, to provide call selection to ensure the achievement of target values for execution metrics based on intervals. Can be used.

The above-mentioned embodiments of the present invention are merely illustrative. For example, in the present invention, the call arrival rate does not need to be regular at predetermined intervals. As will be apparent to those skilled in the art, appropriate adjustments can be made to accommodate various call arrival conditions. It should be understood that the call selection functions referred to herein can be provided by incorporating various device components into the exemplary configuration of the call center shown in FIG. 1. In addition, as mentioned above, the present invention can be applied to various forms of communication other than calls including fax and electronic mail. The contract target value information mentioned above can be determined and executed, for example, administratively or through a computer-telephone integration (CTI) application. As another example, the invention can be implemented in an application programming interface (API) with current call center software packages.

In addition, it should be understood that the present invention, when executed by a computer or other type of processor, may be implemented in the form of a computer-readable medium or other similar medium containing software for causing the processor to execute the above-mentioned processing functions. do. For example, BCMS 110, call vector 140, agent selector 150, and other components in system 101 may be stored in memory 115 or other computer readable medium associated with ACD system 101. One or more software programs stored may each be executed at least in part, and may each be executed by the processor 116 or other processing hardware associated with the ACD system 101. Various other implementations may be used to provide call selection in accordance with the present invention. Various other such alternative embodiments that fall within the scope of the following claims are apparent to those skilled in the art.

In accordance with the present invention, call selection is controlled at the call center to ensure that target values are met for execution metrics based on one or more intervals.

Claims (23)

A selection processing control method for selecting communications for delivery to one of a plurality of agents at a call center, the method comprising: Storing information about performance metrics based on at least one interval for a particular interval; Adjusting the selection process based at least in part on a comparison of the stored value with the value of the metric actually achieved during the interval. 2. The method of claim 1 wherein the execution metric based on the at least one interval comprises at least one of an average rate of the response metric and a percentage of the service level metric. 2. The method according to claim 1, wherein the step of adjusting the service target of the selection process at the designated point within the interval, if the value of the metric actually achieved at the designated point within the interval does not allow achieving the target value within the interval. Selective processing control method comprising the step of changing. The service target of the selection process according to claim 3, wherein the adjusting step comprises: if the value of the metric actually achieved at the designated point in the interval is significantly different from the target value of the metric during the interval, Selective processing control method comprising the step of changing. 2. The method according to claim 1, wherein said adjusting step further comprises determining a fine for out of said stored target value and at least partially adjusting said selection process based on said fine. 2. The method of claim 1, wherein the adjusting step is based on the comparison of the stored target value with a value of a metric actually achieved at a point of a plurality of points during the interval, at the plurality of designated points during the interval. And at least partially adjusting the selection process. 7. The method according to claim 6, wherein the time between at least subset pairs of said specified points is a variable. 8. The method according to claim 7, wherein the time between the designated points is reduced from an initial portion of the interval to a later portion of the interval. 7. The method according to claim 6, wherein the time between each pair of designated points is a predetermined constant. The method according to claim 1, further comprising comparing the stored target value with the value of the metric actually achieved during the interval after each of the plurality of groups for one or more communications have been processed. 2. The method of claim 1 wherein the step of adjusting includes adjusting at least one overload threshold that controls when one or more reservation agents initiate a call. An apparatus for controlling communication selection processing in a call center, A memory for storing information about a target value of an execution metric based on at least one interval for a particular interval; And a processor, coupled with the memory, that at least partially adjusts the selection process based on a comparison of the stored target value with the value of the metric actually achieved during the interval. 13. The apparatus of claim 12, wherein the execution metric based on the at least one interval comprises at least one of an average rate of the response metric and a percentage of the service level metric. 13. The service target of claim 12, wherein the processor is further configured to service the selection process at the designated point in the interval if the value of the metric actually achieved at the designated point in the interval does not allow the goal value to be achieved within the interval. Communication selection processing control unit to change the. 15. The service of claim 14, wherein the processor is further configured to: service the selection process at the designated point in the interval if the value of the metric actually achieved at the designated point in the interval differs significantly from the target value of the metric during the interval. Communication selection processing control device for changing the target. 13. The apparatus of claim 12, wherein the processor is further to determine a fine for being outside the stored target value and to at least partially adjust the selection process based on the fine. The processor of claim 12, wherein the processor is further configured to select the plurality of designated points during the interval based on a comparison of the stored target value with a value of a metric actually achieved at a predetermined point during the interval. A communication selection process control device that at least partially coordinates the process. 18. The apparatus of claim 17, wherein the time between pairs of the subset of designated points is a variable. 19. The apparatus of claim 18, wherein the time between the designated points decreases from an initial portion of the interval to a later portion of the interval. 18. The apparatus of claim 17, wherein the time between each pair of designated points is a predetermined constant. 13. The apparatus of claim 12, wherein the processor further compares the stored target value with the value of the metric actually achieved during the interval after each of the plurality of groups of one or more communications have been processed. 13. The apparatus of claim 12, wherein the processor further adjusts at least one overload threshold that controls when one or more reservation agents initiate a call. When implemented in a processor, the processor, Storing information about a target value of a performance metric based on at least one interval for a particular interval; And call center software to perform at least partially adjusting a communication selection process based on a comparison of the stored target value with a value of a metric actually achieved during the interval.